Identifying Drought-Responsive Crucial Gene In Sorghum By Integrating Gene Differential Co-Expression Networks And Hidden Markov Random Field Model

With the aggravation of global climate change and water supply-demand contradiction,drought has become the main abiotic stress affecting global food production.The growth changes and response mechanisms of crops under drought stress have always been a hot and challenging scientific research topic.Sorghum is a typical drought-resistant crop that can obtain high yield and economic benefits,with low water consumption and high water utilization.Thus,Sorghum is an ideal organism for studying the response of crops to drought stress.However,research on Sorghum transcriptome data in response to drought stress is relatively few.From a data-driven perspective,this paper exploit Sorghum transcriptome with gene differential co-expression network and hidden Markov random field model to mining crucial genes,that responsive to drought stress.The main contents of this paper are as follows.(1)Based on the perspective of gray systems,the gray correlation coefficients are used to measure the correlation between nodes,a novel method for constructing a gene differential co-expression network is proposed.(2)The weighted gene differential co-expression network is used as prior information that is integrated into a hidden Markov random field model.Crucial nodes are screened based on posterior probabilities.(3)The proposed method is applied to the RNA-seq data of Sorghum,crucial genes responsive to drought are mined of each stage.The results are verified by enrichment analysis and sequence alignment,and compared with other methods.The main contributions of this work are as follows:(1)Effective integration of the structure of networks with the model of statistics,we present a method for the data of small sample size to identify crucial nodes in the network.(2)Combining theory with application,crucial genes that responsive to drought were identified of Sorghum at different growth stages,a crucial gene regulatory module is established,and a multilayer differential co-expression network for screened genes reveals that Sorghum is very sensitive to pre-flowering drought.This study systematically explores the key genes of drought response in Sorghum at the transcriptome level,and provides a theoretical basis for exploring the mechanism of drought response in different growth stages of Sorghum,which has potential applications in molecular breeding of drought-resistant crops.